Healthcare organizations are rich in data about care quality and outcomes, but lack generalizable strategies for putting their data to work to improve performance. Giving clinical performance feedback to healthcare professionals is a widely used performance improvement strategy, but evidence about its use shows a pattern of mixed effects over decades of trials. Psychological theory is underutilized in the design of clinical performance feedback, yet it offers robust explanatory mechanisms to improve the cognitive processing and impact of feedback messages. A knowledge-based message tailoring system could reason with theoretical knowledge and clinical performance data to predict optimal feedback message formats and content, while offering explanations of the message design rationale. The research goal of this proposal is to develop and evaluate a knowledge-based message tailoring system for clinical performance feedback. The proposed work will be carried out in the health domain of antimicrobial stewardship, a well-defined domain of global importance in which feedback to clinicians is routinely used to promote behavior change.
The specific aims of the proposed project are 1) Develop a knowledge base for theory-based message tailoring of performance feedback, 2) Create a message tailoring system for antimicrobial stewardship, and 3) Evaluate the function of the prototype message tailoring system with healthcare professionals. By achieving these aims, the candidate will gain research experience and enhance his knowledge about the development and evaluation of knowledge-based systems in clinical settings. The training opportunities created by the NLM K01 award will enable the candidate to enhance his knowledge in ontology development, knowledge engineering, and cognitive studies, and to develop collaborations in the research community at the University of Michigan. The award will ultimately help the candidate to achieve his long-term goal of transforming existing knowledge about message tailoring into computable forms for the purpose of conducting research about the effectiveness of clinical performance feedback and other forms of clinical advice.
Healthcare professionals often receive inadequate feedback about their practice. We propose to build and evaluate intelligent software that uses psychological theory to customize feedback for individuals based on their characteristics and practice history, to increase the relevance of feedback for improving care. The successful completion of this project and future studies will lead healthcare organizations to make better use their data as performance feedback for improving health.
